Seventeenth Annual Meeting. 99 



MM' arc blocks of magnesium carbonate, of a pure white color. The light from -1/ 

 passes through the movable nicol X, the position of which is indicated by an index 

 upon a divided circle, thence through the stationary nicol W into the upper half of the 

 slit. Thus the brightness of the direct spectrum is under the control of the operator. 



The light from M / passes through the colored medium I), and striking the back of 

 the right-angled prism I\ is totally reflected into the lower half of the slit of the spec- 

 troscope. At the focus of the spectroscope a narrow vertical slit is placed, through 

 which hut a small portion of each spectrum is visible. These portions can he compared 

 and made equal in intensity by turning the movable nicol X Thus we can find the rel- 

 ative amount of light of each wave-length transmitted by the medium D. 



< >t" course the movable prism N may be used at the eyepiece with convenience, but 

 in the ease of permanganate of potassium, the light from M' seemed to be partially po- 

 larized by its passage through IK and through the dispersing prism of the spectroscope. 

 This was so marked in one case, that when observing the two spectra, between F and G 

 with the movable nicol at the eyepiece, both would appear and vanish together. 



The curves exhibited in the plate, (fig. 2,) show the light transmitted by solutionsof per- 

 manganate of potassium, of 1 to 1,000, 1 to 5,000, and 1 to 10,000; each having a thick- 

 ness of 1 cm., determined by the method described above. The shaded spectrum, given 

 below, is taken from "The Nature of Light," by Dr. Eugene Lommel, page 175. It 

 shows the usual method of mapping the absorption spectrum of permanganate of potas- 

 sium. One can see at a glance that such maps give a very imperfect idea of the real 

 nature of absorption spectra. 



It will be seen by the above curves that a solution of permanganate of potassium is 

 nearly opaque to yellow- green rays, while the amount of red rays transmitted by con- 

 centrated solutions is much less than that transmitted by dilute solutions. Violet light, 

 however, seems to be transmitted equally well by all solutions. This gives dilute solu- 

 tions a decided red color, while concentrated solutions approach a pure violet. It seems 

 by this that yellow-green rays are reflected, red absorbed, and violet transmitted; but the 

 reflection spectrum has not yet been measured, and to the eye it appears almost white. 



It is my intention to study the absorption and reflection spectra of a series of com- 

 pounds, for instance the chromates, permanganates, ferrates, etc., and also the salts of the 

 metals themselves, to find if there is such relation between them as will furnish a clue 

 to the cause of selective absorption and reflection. 



One cannot study color without wondering why or how a body can show preference 

 to wave-lengths in reflection and transmission. Why a body should reflect violet, for 

 instance, in preference to red or yellow, is an open question. 



The phenomena of opacity and transparency, and the property of absorbing all and 

 reflecting all wave-lengths, as lamp-black and silver are capable of doing, are doubtless 

 closely related to the subject of selective absorption. What molecular relations can 

 there be between black and white, and opaque and transparent bodies? It is surely 

 only a molecular difference. Can the molecules of lamp-black be so loosely hung that 

 they sway under the impingement of any wave-length of light, while those of silver 

 are almost immovable? Another thing, not less remarkable than this, is the color rela- 

 tions between t lie compounds of many of the elements; for instance, many of the metals 

 of the iron group have highly-colored salts. Ferrates, chromates, etc., are particularly 

 remarkable for their color, as are also the salts of the aniline bases. 



We can understand selective reflection only by supposing that the vibrations of the 

 molecules of some bodies deaden the wave-length in unison witli them, while other 

 wave-lengths may he cither reflected or transmitted. This can be readily illustrated by 

 sound. Suppose sound waves of different lengths to strike a screen made of wires stretched 

 very close together and all in unison, then the wave of sound in unison with them would 



